One of the conundrums teams are often challenged with when developing a project schedule is estimating how much time and cost to allocate to tasks when there is a little or no subject matter expertise: How do you estimate something you’ve never done before? In the absence of a crystal ball, here are three techniques to help you estimate the unknown:
Break it down
Don’t try to put a number against something big. It’s difficult to estimate how long it takes to circumnavigate the globe; however, most of us have an idea how long it takes to drive across our hometown; and we can estimate, with a fair degree of accuracy, how long it takes to drive across a town that’s similar in size to ours. Rather than trying to produce one large estimate, try generating a bunch of smaller estimates and then add them together.
Qualify your estimate with assumptions
In the above example, I noted most of us know how long it takes to drive across our hometown. While this may be true, any estimate to do so includes inherent assumptions. For example, our estimate may include an assumption that there is no construction along the route we will take. It may also assume favorable weather conditions. If either of these assumptions proves false, the estimate we provide is likely to be impacted. For this reason, it is critical to include assumptions with your estimate. In doing so, you can begin to manage your stakeholders’ expectations: “Assuming there is no construction along the route and the roads are dry, it will take 90 minutes to drive across town.” Assumptions become risks that you can use to continuously manage stakeholders’ expectations throughout the project so there are no surprises: “Just letting you know, I am halfway across town, now; and there is a snow squall approaching; I will probably be 20 minutes later than planned.”
Resist the temptation to guess
“Guesstimates” will come back to haunt you. Yet, there are times when you simply don’t have enough experience or
information to generate an estimate for even the smallest components or conceive of assumptions. For example, what if you have never driven across your hometown because you have just moved there? How could you estimate the time it ould take for a trip across town (GPS apps aside)? Rather than guessing, you can use a standard deviation (a.k.a. a bell curve) approach to assign probability to a range of estimates. You do this by considering the best case,
worst case and most-likely estimates for your task in order to generate several estimates with varying degrees of accuracy*: “Assuming no construction and fair weather, I should be able to drive across town in 90-100 minutes; however, if these assumptions prove false, it could take 120-140 minutes. On the other hand, if opportunities to pick up some time present themselves, and I can exploit them, my trip across town could take as little as 75 or even 50 minutes.”
* To save you the math, there is a calculator for this at http://www.psplus.ca/taskcal.php
Next time you are asked to provide an estimate for the unknown, don’t guess; don’t cross your fingers and hope for the best. That’s not estimating, that’s wishing. Instead, break large tasks down into smaller ones than can be estimated, qualify your estimate with assumptions, and use a bell curve to generate a range of estimates.
Find more tips like this one at Professional Services Plus